Advancing LLM Reasoning for Theorem Proving Through Natural Language and Reinforcement Learning - View it on GitHub